Kenshi has a generous approach for delivering randomness to your smart contracts. In case of failure we retry twice within 2 minutes of the first try, if all fails we log an entry for it in our db and queue it for a retry in 15 minutes, if that also fails we reschedule again for a retry in 1 hour, then 2 hours, 4 hours, 8 hours, 24 hours and finally 48 hours until we give up.
That's a total of 10 retries. Failure can happen for variety of reasons, for example if you don't have enough Kenshi tokens stored in your smart contract for paying the delivery fees or when the network is down. If the failure is from our side, we will resolve the issues as soon as possible and deliver the results out of schedule.
We use multiple RPC endpoints for delivering your results. These endpoints are chosen based on their uptime and latency. To check the status of the RPC endpoints or average VRF delivery times you can check our status page.